Beyond Analysis and Travelport AI and Machine Learning Teams at the Travix Hackathon
30 Oct 2018
Last week, Travix, the global online travel agency with operations in 39 countries, hosted a Hackathon at their HQ in Amsterdam.
10 teams competed to identify concepts that could be applied to four key themes:
Pimp my journey
World of AI & ML
Fix my foundations
The History Book.
Beyond Analysis competed jointly with Travelport, with two of our data scientists making the trip over. They chose to focus on the World of AI and Machine Learning, looking to solve the blindspot that online travel agents face when customers do not have a digital footprint.
Using clustering methodologies, machine learning and ARIMA models the team created 4 clusters that identified distinct patterns of behavior, as well as predicting flight pricing to a confident degree of accuracy.
Travix are now able to offer their customers a much more relevant product as well as pricing confidence when they book, thereby enabling a significantly enhanced customer experience. At the point of purchase for example, Travix can now let their customers know:
This is the right time to book as this is the best price for this particular itinerary
Wait to be alerted, because prices will decrease
Be given a different itinerary, based on their cluster group, because the prices are above average and there is more value in a different route.
All the teams, made up primarily of internal Travix staff, then reconvened at the end to present back their findings which was a great opportunity to hear about the different approaches used and results achieved.